Warning of lane departure dangers in the safe operation of RTG crane using LaneNet deep learning neural network

Rubber Tyred Gantry Cranes (RTG) frequently experiences lane deviation incidents that disrupt cargo transportations. This paper proposes a lane deviation warning solution utilizing deep learning (DL) neural networks to enhance operational safety. Using actual operating data from RTG cranes, the authors developed the LaneNet neural network to create a warning model that achieves an accuracy of up to 96.3% and a processing speed of 65 FPS. The experimental results demonstrate that this solution significantly improves safety levels while also optimizing the operational performance of RTG cranes. Therefore, future studies on the safety warning system for RTG crane operations will focus on conducting test cases under rigorous conditions using advanced neural networks.